Power converter topologies with series-connected devices must maintain repeatable blocking voltages across the devices to ensure reliability. Effective voltage sharing requires consistent diode reverse recovery charges, which is not guaranteed with commercially produced devices. To evaluate this effect, this paper measures and statistically quantifies the variation in reverse recovery characteristics of diodes from one manufacturer. Statistics are provided for various groupings by date code, shipment tube, and back-and front-end epitaxy processes. The devices are then matched for reverse recovery charge and series-connected in a medium-voltage five-level power converter. Voltage sharing is significantly better than a similar test with devices chosen at random. However, the device voltage mismatch exceeds predictions based solely on charge mismatch.